Please enable JavaScript.
Coggle requires JavaScript to display documents.
DataRobot (Chapter 14: Feature Understanding Selection (data types (binary…
DataRobot
Chapter 14: Feature Understanding Selection
interpret data contained in each feature
using index headers
shows unique items, missing data, mean, stnd dev, median, min, max
physically demonstrated with the use of bar charts showing frequencies
visualization
"bracket"
denotes an inclusive range
parenthesis
exclusive range
displays also via histograms
data types
binary categorical
multi-class categorical
numeric
Boolean
**will tag based on symbols, (ex. $)
Evaluations
"ReferenceID" means it will not be used to predict your target
Missing Values
coded by a '?'
will automatically be replaced with missing values so algorithms can continue to run
Chpt- 15 Build Candidate Models
Starting the process
select the target
detects accuracy
also contains 'advanced options'
"AutoPilot"
Step 1. Setting target feature
Step 2. Creating CV and Holdout partitions
Step 3. Characterizing target variable
Step 4. Loading dataset and preparing data
Step 5. Saving target and partitioning information
Step 6. importance scores
Step 7. Calculating list of models
Chpt 13- Start-Up Processes
easiest to read in a local file
and smaller files are much easier to work with
be careful with CSV files
store data as a .tsv or.xlsx
begin a new file "Untitled Project", new projects can be created here, use manager projects
"manage projects lists all prior and currently running projects"
use tags
exit projects
"clicking the data link"